Alexandria Catalogue - Highlights

Memory Integrity Enforcement (MIE) is the culmination of an unprecedented design and engineering effort, spanning half a decade, that combines the unique strengths of Apple silicon hardware with our advanced operating system security to provide industry-first, always-on memory safety protection across our devices — without compromising our best-in-class device performance. We believe Memory Integrity Enforcement represents the most significant upgrade to memory safety in the history of consumer operating systems.

But that’s a minor problem for some $WLFI buyers. Hours after the market closed, some buyers began to report on the official WLFI forum that they were being targeted by hackers who were draining their accounts of new coins. $WLFI operates on the ethereum blockchain, which had a recent upgrade that hackers used to exploit the accounts to steal their coins. “The instant the tokens arrive, they will be stolen by automated sweeper bots before we have a chance to move them to a secure wallet,” one user wrote.

/sigh don't believe spam belongs on the bitcoin network. ofc, no one does! it's just filtering does not work.

Substack writers just saw their prices inflated 30% on iOS. Writers now have to wait 45 days to receive payment from Apple. Writers had absolutely no say in the decision, despite the very real consequences for their business. The feature cannot be disabled or toggled off.

Stop flushing your performance down the drain.

This is overtly political and has nothing to do with being “nice” or “empathetic”. The corporations have inadvertently made an arrangement with critical theory leftists to transform society into a capitalist monopoly on top, and socialism on the bottom. That’s the deal. It’s been dubbed “Woke Capitalism”, where the managerial class has reached a compromise in which they get to keep and - through being appointed as stewards of the political movement - gain power, while activists of the movement get the transformation they demand. This particular warning in Google Docs is just that. It’s Google managing the changes in society that certain activists want, in order to maintain corporate power. Changes like this will keep being added (as we should know by now) while our social structures change into something comparable to a social credit system. These changes are promoted by manufactured social pressures and financial incentives from large investment banks and so are implemented by any company that wishes to grow while under public scrutiny. Discussions about this have been like groundhog day since 2015, no one seems to be able to get past the surface level of what is probably the greatest transformation of any society since the 1930’s, and it’s astonishing. It’s a gradual escalation of demands from woke activists, backed by the behemoth corporations and banks, then adopted by governments. For anyone who wants to help others, or to resolve injustices in our society, having true empathy for the plight of those oppressed and wanting to solve social problems: you know this is bullshit. Even for the activists this isn’t about actually helping anyone, it’s about pushing us into accepting a shuffling of hierarchies, of which they didn’t bother to iron out the injustices. https://news.ycombinator.com/item?id=31090216

As always, you should continue sharing Spotify content directly through your favorite platforms like Instagram, Facebook, WhatsApp, Snapchat, TikTok, and more. Messages on Spotify are designed to complement these integrations, not replace them, and we’re excited to continue offering more ways to drive hype for the Spotify content you love, wherever you are.

In a world pushing for reflexive AI usage, senior developers who practice intentional collaboration (if any!) with AI while preserving the craft of programming become the guardians of software quality. They’re the ones who ensure that our systems remain comprehensible to humans, true to their domains, and built upon solid theoretical foundations.

Naur’s essay reminds us that programming is fundamentally a human intellectual activity. The real product of programming isn’t the code—it’s the theory, the understanding, the mental model that gives the system coherence. While LLMs can generate syntactically correct code, they cannot participate in theory building. They cannot understand business context, make thoughtful trade-offs, or maintain the conceptual integrity that separates good software from mere working code.

Domain-Blind Code Generation represents an even more extreme case. LLM-generated code isn’t just theory-less—it’s nobody’s theory. It emerges from statistical patterns without understanding the business domain, the system’s conceptual model, or the nuanced trade-offs that domain experts have encoded into the architecture. The code might pass tests, but it exists in a theoretical and domain vacuum. With no git blame pointing to a human being at a nearby desk, I’m at a severe disadvantage trying to discern the how and more importantly the why of the code.

Naur argues that programming is fundamentally about building a theory—a shared mental model of how a system works, why it works that way, and how it should evolve. The source code is merely a written representation of this theory, and like all representations, it’s lossy. Critical knowledge about intent, design decisions, trade-offs, and the reasoning behind architectural choices exists only in the minds of the people who built the system. When those people leave, the theory dies with them. The code remains, but the program—the living, breathing system with its context and rationale—is gone.

Knots has code to turn itself off at a set expiry date. Unnecessary, stupid, and outright dangerous. Did you know? Was it discussed in a team of experts? How can such a dangerous "feature" even make it into a release? By trusting a single guy who lives in his own reality.

Markdown parsers are supposed to leave HTML tags untouched

Iterate Fearlessly When you encounter errors, don't give up. Just say "retry." Embrace the iterative process of prompting, observing, refining, and repeating. Every failure is a learning opportunity, not a dead end, so keep experimenting and adjusting.

But that's not the only way in which an LLM coding assistant is like a slot machine. Reg Braithwaite proposed that AI companies' business model is also like a casino's, because they charge every time you re-prompt the AI. He writes: When you are paying by the "pull of the handle," the vendor's incentive is not to solve your problem with a single pull, but to give the appearance of progress towards solving your problem.

If Bluesky wants to protect itself from legal liability, let it do what every company did until just a couple years ago: a) don't break the law on purpose, and; b) buy insurance These new ToS are an absolute dog's breakfast. I wouldn't click through them.

As things stand, Bluesky has very bad terms of service that every user who creates an account has to subject themselves to. In particular, Bluesky's ToS contain a "binding arbitration" waiver that forces users to surrender the right to sue Bluesky no matter how the company harms them. This is so pro-enshittificatory, it's like a landing strip for the sole use of Enshittification Airlines, which can land a 747 full of enshittfying nonsense on Bluesky's users every 10 minutes, around the clock, without worrying about any legal repercussions.

By moving the Podcasting 2.0 payments standard to thelnaddressformat, we can simplify the user experience for everyone involved, and give app developers an easy way to let listeners connect their favourite wallet

It’s time to evolve from trusting the transport to securing the content.

Prime Minister Benjamin Netanyahu tells i24 he feels he is on a “historic and spiritual mission,” and that he is “very” attached to the vision of a Greater Israel.

An engineered protein that acts like a molecular sponge has the potential to change how carbon monoxide poisoning is treated, chasing down CO molecules in the bloodstream and helping the body flush them out in just minutes, without the risk of short- or long-term health issues that come with the current frontline treatment, pure oxygen.

The value of technology comes from effectively solving problems for people. By that metric, many simple technical solutions are more valuable than complex ones.

we don't need a return to overwhelming engagement or a descent into complete detachment, we need a careful curation of where we invest our emotional energy. I don’t think it’s a binary choice between feeling everything or nothing - there is (there has to be) a spectrum of feeling the right things deeply enough to act on them.

"The mistake is thinking we need to feel everything or nothing. The skill is choosing what to care about and accepting that choice as enough."

GPT-5 is here. Rolling out to everyone starting today.

This proposal introduces a self-custodial credit system that sits in front of an existing ecash mint. The users balance is held in a trustless credit contract that can be instantly swapped for ecash tokens offchain at the time of payment. Custodial risk only exists while payments are in flight, which is usually seconds, and only for the amount currently being transacted. The balance is always self-custodial and if the mint disappears or is uncooperative the user can unilaterally exit and reclaim their credits onchain.

If you're skeptical about Bitcoin being the foundation of the world's new monetary and financial system, remember Linux was once a mere hobby project and is now the foundation of the Internet and modern computing as we know it.

When Perplexity's official crawlers got blocked, the company just deployed masqueraded user agents pretending to be regular Chrome browsers on Mac computers. These undeclared crawlers used rotating IPs and different ASN attempts to further dodge website blocks.

I suspect many of us have had the experience of not using a skill long enough to completely un-learn it. LLMs enable this, but with every skill.

“LLMs are doing reasoning” is the “look, my dog is smiling” of technology

Git is a decentralized version control system, yet most freedom tech projects use centralized walled gardens on top of git as a social and collaboration layer for code changes.

a package called google-chrome-stable doesn’t just install Google’s browser, but also runs a RAT on your system. That kind of malware can potentially give attackers control over the infected machine, letting them steal data, install more malicious software, or spy on users.

Two potatoes are two potatoes. Ten thousand bricks are ten thousand bricks. Potatoes and bricks don't exhibit emergent properties as they come together. People do. The meaning of human in human-centred must be read therefore as being human in relationships, in communities, and most definitely not individual or user or consumer.

In a time of rising inflation and centralised control, Bitcoin offers a peaceful alternative. It is open, transparent, and accessible.

Building something real means making difficult trade-offs in a world of constraints. It is not about building something that will be perfect forever. It's about building something that provides a real benefit to real people; something that puts the user's sovereignty first; something that empowers the individual, instead of trapping them or selling them.

true understanding means explaining concepts in your own words and contextualizing them for your specific audience

SSI has been described to me as “the most contextually-sensitive identity technology I know of” in contrast to my assertion that SSI cannot communicate context (Sheldrake 2021c). You won’t be surprised to learn that we’re both right in our different contexts. The first context is the rather low bar of identity technology, whereas mine is human society and nature, and what SSI sets out to do contextually is an utterly miserable substitute for contextually-sensitive non-technologized human relations. This may well be a criticism of the general datafication of human society, but SSI is positioned as a flag carrier, as a primary vector in this regard. The technology ignores contexts by design.

The frequency with which people need to trust each other declines as SSI use grows, which in turn changes the character of human relations, of community, of society. Today, we largely qualify each other contextually with a difficult to describe but deeply valuable fuzziness. Tomorrow, we, or more precisely our digital agents, quantify and filter non-contextually with exacting precision, with unforgiving recall, and with mindless programmatic dedication.

Whereas ‘backwards compatibility’ typically conveys interoperability with an older, less capable system, in our context the ‘legacy system’ is of a massively more powerful design. Compatibility may be pursued with attention to contexts, to variety, to restraint, to the interdisciplinary analysis of existing cultural patterns, and to the accretive development of social norms. Ultimately, any sufficiently advanced technology will be indistinguishable from Nature (Schroeder 2011).

The SSI community, emanating from and largely operating with the ALONE paradigm, sets out in good part to remove The State as the issuer of identity, but caught in the binary thinking characteristic of ALONE, replaces it with something less social, less easily governed, less accountable, less easily restrained and more insidious, in stark opposition to the social layers and norms and institutions and mechanisms integral to what might be referred to as high-trust societies and indeed the trust-based essence of all human community (Cook 2001; Fukuyama 1995; O’Hara & Hutton 2004). Renieris (2021) describes the SSI framework as based on the “neoliberal fallacy of individual choice & control.”

Self-sovereign identity (SSI) is premised on simplistic political dogma. Just like all simplistic political dogma, it breaks as soon as sufficient pain is felt by sufficient numbers of people retaining the power to revolt. The SSI community, if I can refer to a group identity, believes that giving Alice her self-sovereignty frees her, when in fact it requires that she conforms to one particular political worldview. The system subsumes her and her networks, both internal and external.

Humans are social animals. Community, and more accurately cooperation, is essential to our surviving and thriving (Margulis & Sagan 1997; Nowak & Highfield 2011). Organising community entails governance at many scales, and we refer to the activities of governance as politics. All sociotechnology — relating to the technologization of human organising — is therefore political.

Plus, thanks to agents getting reasonably good at coding, you don’t have to do the groundwork of sitting down for hours to produce a working prototype. You can literally [speak your ideas into existence](https://github.com/dergigi/vibeline), as we did in the last cohort. And once something exists, you can properly and meaningfully talk about it, and pursue it further—or not.

If Bitcoin doesn’t need a state, its users are still bound to some jurisdictions — but they do have a choice. And increasingly, the choice is: build where the rules work for you, or get buried under rules that don’t.

the advent of LLMs and "vibe-coding" now allows us to enter into an ongoing dialogue with technology itself

Test

The catalyst for this change is the maturation of AI, specifically large language models that can understand instructions, write code, interact with tools, spin up their own temporary computing environments, and then tear everything down the moment a task is complete. This is the “appocalypse”: not a destructive event, but a fundamental transformation from static code to dynamic, agentic systems. In this new world, you do not go to an app to perform a task. You simply state your intent in natural language, and a domain-specific agentic framework assembles the necessary tools and code in real time to deliver the result you need.

Identities are immanent in contextual relationships, and relationships are immanent in information exchange. A person is a person through other persons (Birhane 2017).

Slack is becoming so ran thru. MS Teams is a non-starter. Discord is gamer slop. I just want channels, threads, and DMs. Like nothing else. Why does every product tend towards slop like this?